Developers
Set your API contextPick a region and space ID once. Every sample fills in your base URL.
Where do I find these?

In Coevera: open your space's API application and choose Show Access. It shows the space ID and the base URL, whose host starts with your region.

From the API: if you know the space ID, the discovery endpoint GET /api/discovery/spaces/{space_id} returns its region and base URLs, without credentials.

Only the region and space ID are kept, in this browser. Never paste a key or password here.

More in the quickstart

AI & MCPHosting and your data

Hosting and your data

Where the MCP server runs, where your CRM data stays, which logs it keeps, and the path file uploads and exports take. For IT and security reviews.

Answer

The MCP server is hosted in the United States, whatever the region of your space. Your CRM data stays stored in its region, but requests and answers pass through the US server. Usage logs are kept mostly in Ireland for up to 90 days, with a 24-hour copy in the US, and error reports go to a US-based error-monitoring service.

These are the answers an IT or security review usually asks for first. They describe the server at https://mcp-server.coevera.com, which every space uses.

Where things run

Topic Today
MCP server Hosted in the United States, for every space, whatever its region. There is no regional MCP server.
Your CRM data Stays stored in your space's region: us-east (Virginia), eu-central (Frankfurt), ca-central (Toronto) or ap-southeast (Sydney).
Data in transit Every tool call travels from the assistant to the US server, then to your space's API in its region, and the answer travels back the same way. A space in Frankfurt, Toronto or Sydney therefore sends assistant traffic through the US.
Usage logs Basic usage logs (who made the request, which company account, which request type) are kept mostly in Ireland for up to 90 days, with a 24-hour copy in the US.
Error reports Sent to a US-based error-monitoring service.
Status Beta. The connectors are free.

Sign-in and credentials

  • Users sign in on Coevera's own sign-in page. The assistant receives an access token, never the user's password.
  • The server checks the token on every request; it keeps no session between requests.
  • Each user's access is limited by their own permissions in the space and by the scopes they approved; see OAuth and scopes.
  • Changes made through the connector are recorded in each record's history like any other change.

File uploads

The file upload reference of the MCP server gives both steps in full, with an example. A file upload takes two steps. The upload_cloud_object_document tool (a document) or the set_record_picture tool (the picture of an account, contact or product) returns a single-use upload URL on mcp-server.coevera.com; the assistant's environment then posts the file to that URL, and the server stores it in Coevera under the user's identity.

  • The upload URL works once and expires after five minutes.
  • Files are limited to 50 MiB. A picture must be a PNG, JPEG, GIF or WebP image; the server checks the file's content, not its name.
  • The assistant's environment must be allowed to reach *.coevera.com.

Exports

export_records writes the records that match a filter to a CSV file, also in two steps.

  • The tool checks that the user may read the entity and holds the Export permission in Coevera, counts the rows, and returns a one-time download URL on mcp-server.coevera.com. The response describes the file and its columns; it holds no rows.
  • Opening the URL streams the file straight from your space's API under the user's identity, so record visibility and field permissions apply. The file is never stored on the MCP server, which keeps only the export request, encrypted, until the link is used or expires.
  • The download URL needs no sign-in: whoever holds it can download the file once, until it expires (after 10 minutes by default). Treat it like the file itself.
  • Clients ask the user to approve every export, because the tool is not marked read-only. The assistant may download the file into its own environment to analyse it, or hand the link to the user.

What the assistant does with the data

Tool results go to the assistant you connected. How that provider processes and keeps them is set by the provider's own terms, not by Coevera.

The server's instructions tell the assistant that tool results are data entered by users, never instructions to follow.